Blab AI
Qualcomm to Acquire AI Startup Modular for Nearly $4 B, Expanding Cross-Chip AI Software
On June 24, Qualcomm Incorporated (NASDAQ:QCOM) announced it would acquire AI‑software startup Modular in an all‑stock deal worth almost $4 billion. The transaction will give Qualcomm a platform that lets artificial‑intelligence models run on a variety of processors without developers having to write separate code for each chip.
The deal marks a decisive step for Qualcomm, positioning it to compete with Nvidia’s CUDA platform—a long‑standing industry standard for programming GPUs in AI workloads. CUDA’s ecosystem has attracted a large developer base, and Qualcomm’s new software aims to deliver a comparable level of cross‑platform support for its own chips.
Modular, founded in 2022, has raised $380 million in funding, including a $250 million round in September of that year. Reuters reported that Qualcomm will issue up to 19.2 million shares of its common stock to Modular’s equity holders as part of the transaction.
Qualcomm’s core business centers on designing and marketing semiconductor technologies for mobile devices and other wireless products. The company operates through divisions such as Qualcomm CDMA Technologies, Qualcomm Technology Licensing, and Qualcomm Strategic Initiatives.
The acquisition follows reports that Qualcomm is also in advanced talks to buy AI‑chip startup Tenstorrent for a valuation between $8 billion and $10 billion. According to The Information, the discussions involve a RISC‑V‑based AI accelerator designed by former Intel engineer Jim Keller.
Modular’s software is intended to enable AI models to be deployed across different hardware platforms, reducing the need for platform‑specific optimizations. This capability could streamline development for companies that use Qualcomm’s Snapdragon processors in smartphones, automotive systems, and edge devices.
The move comes amid a broader trend of semiconductor firms expanding into AI infrastructure. Qualcomm’s strategy appears to be two‑fold: strengthen its position in the mobile and embedded AI markets while also targeting data‑center and autonomous‑vehicle segments.
Industry analysts note that the deal could help Qualcomm compete with Nvidia’s entrenched CUDA ecosystem. CUDA has been a key driver of Nvidia’s dominance in AI, providing a unified API that allows developers to harness GPU power for training and inference.
Qualcomm’s acquisition of Modular also aligns with its recent announcements of AI‑centric products, such as the AI200 and AI250 chips designed for rack‑scale data‑center inference. The company has highlighted the importance of software tools that can bridge the gap between different hardware architectures.
While the deal is still subject to regulatory approval, the announcement signals Qualcomm’s commitment to building a comprehensive AI stack that includes both hardware and software components.
The potential Tenstorrent acquisition would further extend Qualcomm’s reach into the AI‑chip market, offering a complementary architecture that could diversify its product portfolio.
In summary, Qualcomm’s purchase of Modular for nearly $4 billion is a strategic step to broaden its AI software capabilities and challenge Nvidia’s CUDA dominance. The company’s ongoing talks to acquire Tenstorrent suggest a continued push into AI‑chip development. The outcomes of these deals will shape the competitive landscape of AI hardware and software over the coming years.
The current status of the Modular acquisition is that Qualcomm has agreed to issue shares to the startup’s equity holders, and the transaction is expected to close after regulatory review. The Tenstorrent talks remain at the negotiation stage, with no definitive timeline announced.
As the semiconductor industry continues to evolve, Qualcomm’s moves into AI software and chip design reflect a broader trend of companies seeking to control both the hardware and software layers of AI deployment.
The deal marks a decisive step for Qualcomm, positioning it to compete with Nvidia’s CUDA platform—a long‑standing industry standard for programming GPUs in AI workloads. CUDA’s ecosystem has attracted a large developer base, and Qualcomm’s new software aims to deliver a comparable level of cross‑platform support for its own chips.
Modular, founded in 2022, has raised $380 million in funding, including a $250 million round in September of that year. Reuters reported that Qualcomm will issue up to 19.2 million shares of its common stock to Modular’s equity holders as part of the transaction.
Qualcomm’s core business centers on designing and marketing semiconductor technologies for mobile devices and other wireless products. The company operates through divisions such as Qualcomm CDMA Technologies, Qualcomm Technology Licensing, and Qualcomm Strategic Initiatives.
The acquisition follows reports that Qualcomm is also in advanced talks to buy AI‑chip startup Tenstorrent for a valuation between $8 billion and $10 billion. According to The Information, the discussions involve a RISC‑V‑based AI accelerator designed by former Intel engineer Jim Keller.
Modular’s software is intended to enable AI models to be deployed across different hardware platforms, reducing the need for platform‑specific optimizations. This capability could streamline development for companies that use Qualcomm’s Snapdragon processors in smartphones, automotive systems, and edge devices.
The move comes amid a broader trend of semiconductor firms expanding into AI infrastructure. Qualcomm’s strategy appears to be two‑fold: strengthen its position in the mobile and embedded AI markets while also targeting data‑center and autonomous‑vehicle segments.
Industry analysts note that the deal could help Qualcomm compete with Nvidia’s entrenched CUDA ecosystem. CUDA has been a key driver of Nvidia’s dominance in AI, providing a unified API that allows developers to harness GPU power for training and inference.
Qualcomm’s acquisition of Modular also aligns with its recent announcements of AI‑centric products, such as the AI200 and AI250 chips designed for rack‑scale data‑center inference. The company has highlighted the importance of software tools that can bridge the gap between different hardware architectures.
While the deal is still subject to regulatory approval, the announcement signals Qualcomm’s commitment to building a comprehensive AI stack that includes both hardware and software components.
The potential Tenstorrent acquisition would further extend Qualcomm’s reach into the AI‑chip market, offering a complementary architecture that could diversify its product portfolio.
In summary, Qualcomm’s purchase of Modular for nearly $4 billion is a strategic step to broaden its AI software capabilities and challenge Nvidia’s CUDA dominance. The company’s ongoing talks to acquire Tenstorrent suggest a continued push into AI‑chip development. The outcomes of these deals will shape the competitive landscape of AI hardware and software over the coming years.
The current status of the Modular acquisition is that Qualcomm has agreed to issue shares to the startup’s equity holders, and the transaction is expected to close after regulatory review. The Tenstorrent talks remain at the negotiation stage, with no definitive timeline announced.
As the semiconductor industry continues to evolve, Qualcomm’s moves into AI software and chip design reflect a broader trend of companies seeking to control both the hardware and software layers of AI deployment.
